Blog

How to build a competitor playbook with an AI agent, for 11 cents

A competitor playbook usually takes a product marketer a week. Read the pricing page, skim G2, search Reddit, check the ad libraries, look at their careers page, read the changelog, then write it up for sales. By the time it's done, half of it is out of date.

I asked an agent to do it on November 9, as if I were selling a project management tool against Linear. It bought everything it needed in about eight minutes, used nine services and spent at most $0.113. Then it wrote a playbook my sales team could use on a call.

Part of the playbookWhat the agent readWhat it cost
PricingLinear's pricing and billing pages$0.02 for reading pages
Complaints10 G2 reviews, 30 Reddit posts, 16 comments$0.044 for G2, $0.00018 for Reddit
AdsMeta and Google ad libraries$0.021, nothing usable
HiringLinear's careers page, 24 rolespart of web search
NewsChangelog and coverage, last 30 days$0.00275

By hand

  1. Read the pricing page
  2. Skim G2 reviews
  3. Search Reddit
  4. Check two ad libraries
  5. Read the careers page and changelog
  6. Write it up for sales
  7. about a week

With an agent

  1. Ask for the playbook
  2. Pick how much Reddit to read
  3. Read the playbook
  4. about 8 minutes, $0.113
The same playbook, by hand and with an agent.

Step 1: Ask for the whole playbook at once

In Delegate Agents:

Build me a competitor playbook on Linear, for a team that sells a project management tool. Include what Linear charges, what its users complain about on Reddit and G2, which ads it runs on Meta and Google, which roles it is hiring for, and its news from the last 30 days. A source for every point. Keep the whole job under $3.

Say who you are in the prompt. "For a team that sells a project management tool" is what turned a research summary into discovery questions and objection handling.

Step 2: Answer its question

Before buying anything it asked how much of Reddit to read, about 30 posts or about 100. I took the recommended 30.

Step 3: Check the pricing it verified

It read Linear's pricing page and put the plans in a table: Free, Basic at $10 a user a month, Business at $16, Enterprise on request, each with its limits. It also flagged what it could not find. Monthly-billed prices weren't on the page, so it didn't guess them.

The start of the playbook in Delegate Agents: an executive take with the recommended competitive angle, a coverage note saying Meta and Google ad creatives could not be verified, and a table of Linear's Free, Basic, Business and Enterprise plans with prices and limits.
The executive take and the pricing table. Delegate Agents, November 9, 2026.

Step 4: Turn the complaints into questions

This is where the agent was better than most playbooks I've seen. It read ten G2 reviews and 30 Reddit posts and found real friction: noisy notifications, reporting that needs setup, permissions that confuse new team owners, a learning curve on bigger workflows.

Then it turned each one into a discovery question instead of a claim. "How do leaders get a progress view today, and how much configuration does it take?" It also told me what not to say. Reviewers praised Linear's speed and design over and over, so "Linear is too hard for non-engineers" is "not a defensible blanket claim."

Step 5: Notice what it couldn't verify

It searched the Meta and Google ad libraries and found only unrelated advertisers with similar names. So the playbook says it could not verify Linear's ads, and that "unverified does not mean no ads are running." No invented ad themes, no made-up spend. The two ad searches still cost two cents. (A human analyst under deadline would have been tempted.)

Step 6: Use the hiring and news to stay current

Linear's careers page listed 24 roles, many of them sales roles in Singapore, London and North America, next to AI, mobile and security engineering. The agent read that as international expansion and continued enterprise investment, and labeled it as its own interpretation. From the changelog it picked three dated updates from the last 30 days, each with what it means for a sales call.

Step 7: Take the objection handling into the call

The last section is the one sales will use: which deals to prioritize, a demo to run on one real project, and answers to the four objections you'll hear.

The end of the playbook: a demonstration to run, answers to four likely objections such as Linear is faster and Our engineers love it, a suggested opener, and the receipt line Used 9 services, up to $0.112755.
The demo, objection handling and opener, with the receipt at the bottom.
services used
9
G2 reviews, Reddit posts and comments read
56
to gather everything
about 8 min
at most, for the whole playbook
$0.113
The run in Delegate Agents on November 9, 2026.
G2 reviews
$0.044
Web search, 6 calls
$0.025
Reading web pages
$0.02
Google ad library
$0.018
Meta ad library
$0.003
Company news
$0.003
Reddit, 3 calls
$0.001
What the playbook cost, by service. $0.113 in all.

Make it a monthly habit

Pricing pages change and changelogs move fast. Any chat in Delegate Agents can repeat on a schedule, so the same prompt can run on the first of every month and tell you only what changed.

Questions people ask

Can I run it on several competitors?

Yes. Ask for the same playbook on each, or one comparison across three. Cost grows with how many reviews and posts it reads, so keep the "under $3" line in the prompt.

What if the competitor runs ads I know about?

Say so in the prompt, with the advertiser name as it appears in the ad library. Same-name companies are the usual reason an ad search comes back empty.

Will it just repeat their marketing?

It quotes their positioning as positioning, and marks its own reading of the evidence as interpretation. Check the sources it links before you put a claim in front of a buyer.

Open Delegate Agents and paste the prompt with your competitor and what you sell. Which competitor would you run first?

Give your agent the whole catalog.

Get started

No subscription needed. You pay per job.